Field-crop transcriptome models are enhanced by measurements in systematically controlled environments

Plants in the field respond to seasonal and diel changes in various environmental factors such as irradiance and temperature. We previously developed a statistical model that predicts rice gene expression from the meteorological data and identified the environmental factors regulating each gene. However, since irradiance and temperature-the two most critical environmental factors-are correlated in the field, it remains difficult to distinguish their roles in gene expression regulation.

We show that transcriptome dynamics in the field are predominantly regulated by irradiance, by the modeling involving diel transcriptome data from the 73 controlled conditions where irradiance and temperature are independently varied. The model's prediction performance is substantially high when trained using field and controlled conditions data.

Our results highlight the utility of a systematic sampling approach under controlled environments to understand the mechanism of plant environmental response and to improve transcriptome prediction under field environments.